A modem that offers a connection to another project is found by six
characters derived from its project and itself (modemCode): a copy of the
project has another code, so no two modems offer the same one. The room
names its modems of other projects to F's relay at a start and none at a
stop; the relay pairs each offer with the first modem that joins its code,
one to one as Java's server modem takes one client, and passes the frames
between the rooms' threads. Every client hears how each modem stands
(`modems`: connected, waiting, taken, unpaired, ambiguous, detached, or
unavailable without F). The document keeps the connection and the code.
Measured on F with two threads: a ping from one project's notebook to the
other's goes through; a third project that wants the same code reads
taken, and takes it when the second stops; a deleted project's modem
leaves with its room.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Measured first: tools/java-oracle's ModemRun joins two networks by a pair
of Java modems over TCP on localhost. The modem has no card and adds no
delay of its own — a frame takes the cables on both sides, nothing more.
The core pairs the two enabled modems on a port (Swift's inProject, with
its conditions and cycle suppression) on Java's timing: no 20 ms. The
modem example of FILIUS now pings through instead of failing the core.
A modem of another project (relay) gives its frames to whoever runs the
engine and puts on its cable what comes back, with this engine's frame
and packet numbers; a frame dies after MAXIMUM_RELAY_HOPS trips.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Beside the dashboard only: Java's import window gets "Aus meinen Dateien"
(D4) — the dashboard is asked over the bridge for a file of at most Java's
150 000 bytes and hands its bytes over — and a file's menu gets "In meine
Dateien speichern" (D5), its bytes handed to the dashboard to save.
- Bridge: chooseFile/chosenFile and saveFile/savedFile, each question with
its number; an answer from elsewhere, of another version, for no question
or of the other kind is no answer.
- Whoever reads or finds the project frozen may save, not bring in.
- doc fileBytes: a text as UTF-8, a picture as it is, and "missing" while
a blob's bytes have not come (read so, the file would be empty).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New oracle DocuRun on Java's JDocuElement. Italic is kept (the model knew
only bold), an item keeps its file's size and font size (Swift's 32 x 24
and 8-72 changed Java's files), an item without a colour is Java's light
grey. Java's new items, colour lists, font sizes and styles, the drag's
smallest sizes and Java's stacking; document commands for the editor.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
One row per frame (its highest layer's), Java's detail tree, the table as
Java's text file, "Tabellen löschen" for every card (clearCapture in the
room), a tab per card with Java's title, Java's layer colours as tokens,
the chosen frame and autoscroll per tab. New oracle ExchangeWindowRun.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ViewersRun takes the rows of Java's viewers from the firmware: a switch's
table, a home router's NAT table, each after its reset, and the frames
"Pakete verwerfen" drops while it is held. The core writes a frame as Java
does and resets the NAT's dynamic entries; the room sends both tables as
views, drops while any client holds the button, and sends the dropped
frames, begun anew with each start, with the clock their times of day are
written from; the editor opens a table with a tap on a switch or home
router and holds "Pakete verwerfen" in its toolbar, which now wraps on a
phone. This completes F4.6.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GnutellaWindowRun uses Java's Gnutella window on terminal-net beside the
line of GnutellaRun's story "search". The core cancels a search as Java
does (every connection of the program's own requests closed, passed-on
queries kept), lets the number of peers change and lists /peer2peer's
files; the room takes the window's actions and sends its view; the editor
shows Java's four tabs with Java's symbols. Not adopted from Java: a number
of peers of one digit only, a results table that keeps the last search's
rows, a files table that lags behind the folder.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The firewall-pc scenario and FirewallWindowRun use Java's window as a person
does and send a ping, a TCP connection and a DNS question from the server's
network and from behind the router after each step. The core adds and
removes port rules (personalFirewall.ts) and writes both into the firewall's
log, which now carries each line's time; the room takes the window's
actions and sends its view and the log as lines; the editor shows Java's
window with its two tabs. Not adopted from Java: any integer as a port,
a port after a space refused without a word, a log one can type into.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Oracle DnsWindowRun uses Java's DNS server window as a person does and
asks a computer's resolver after each step. The core's DNS server is a
program its window changes: on and off (its socket on port 53),
recursive resolution, and records that are /dns/hosts, read on every
query as Java's server reads them (the reader moves from fls to the
core). The snapshot carries every installed DNS server with whether it
runs. The room takes the window's actions, sends its view and writes
the server back after the run; the editor shows Java's window with its
three tabs. Not taken from Java: a cell edit spilling into records of
another type, an add losing what another program wrote into the file,
an empty mail domain, a removal hitting the first record of the name.
A table cell no longer takes the room's value over what is being typed
(the web server's too).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Oracle EchoWindowRun uses Java's echo server and client windows as a
person does; SaveAfterRunRun echo saves what Java keeps after the run.
The room takes setActive with a port, connect/disconnect/send, sends
both views and writes the server's port and the client's target back.
The editor's windows show Java's buttons, fields and lines; a port that
is no port is marked and not handed over (Java's button throws).
Core fix: an empty message (one empty segment with the push, as Java's
erstelleSegmente makes) counted as a repeat and never arrived.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Oracle FileExplorerWindowRun uses Java's explorer and viewer windows as a
person does (the right-click menu, Java's questions answered, the import
window, the viewer's open). Both windows work on the document's files
through the core's operations: the tree of folders and files, the chosen
folder's entries, Java's menu (also as buttons, no right click needed),
copy that stays and cut that goes once pasted, the import; the viewer's
picture and title. Not taken from Java: a file chosen in the tree taken
for a folder, pictures laid over each other, a text file as an empty
picture. writeFiles numbers the entries in the file system's order again
when a rename made a new key.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Oracle TextEditorWindowRun uses Java's text editor window and its file
dialog as a person does. The open file's text is its Y.Text in the
document (plan 7.3): typed is in the file at once, for everyone and for
the running engine. Core openText/saveTextAs without Java's mistakes
(lost saves after save-as onto a file or a removed file, files named
'ordner/d.txt' or ''); doc files.ts (changeFiles, fileText, nodeFiles,
readFilesInto); the room takes a person's files into the engine at once
and writes the engine's files back at the tick they changed; the editor's
window, Java's dialog, the text area following others' changes with the
caret in place, its own undo.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The oracle SaveAfterRunRun saves Java's project after a run. The room writes
what a run leaves that Java keeps - files, running servers, virtual hosts,
mail servers' domains and accounts, mail programs' accounts and folders,
firewalls, DHCP clients' addresses - into the document as origin 'sim', at
most once a second, after each action and at the stop; writeRunState writes
only what differs. The command line now goes on in a folder made before a
restart, as Java's does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The oracle MailWindowRun uses Java's e-mail windows as a person does and
answers their dialogs. The core sets and saves the program's account,
deletes a mail, checks a form as Java's window does, and gives the server
its domain, new and removed accounts and the window's own line; the room
sends both programs' views and takes their actions; the editor has both
windows. Better than Java: a name field for a new account, a user name the
server has is refused, deleting hits the folder shown, a new domain goes
into the accounts file at once.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The oracle BrowserWindowRun uses Java's browser and web server windows as a
person does; the room keeps a browser window per computer and sends a
program window's view, the start names the person's language. The editor
shows the page in a sandboxed frame, links and forms go to the room, the web
server has its button, virtual hosts and lines. Better than Java: relative
images and links count from the page's folder, a mark stays in the page, the
error page from Java's template, the virtual hosts' file with complete rows.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A new oracle, InstallRun, installs each of Java's programs on new
computers, all of them on one, each installed and removed again, and a
DNS and an e-mail server installed again after their removal; Java saves
the result as a project. A new program is switched off (a computer's
firewall is on), makes the folders and files it lacks (the web server's
start page in FILIUS' language and its picture, /peer2peer,
/mailserver/konten.txt, /dns/hosts), and removing it takes its settings
and leaves its files; a server installed again reads its records or
accounts from them. Java installs in the order of its list and starts
the program at once.
The document's setSoftware does the same, checked against Java's saved
installation. The core knows what a program makes in the files and
installs or removes a program while the simulation runs; the room
watches the document's programs and hands changes to the engine, so a
new command line works at once. The desktop offers Java's dialog with
two lists, arrows, double click and apply changes.
Also: Java's file may name a number it rolled by the result of any
getter (an idref, here a thread's priority); GetterDefaults now
measures the priority of the saved thread classes, which the import
read as 0 before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A new oracle, TerminalWindowRun, types into Java's own command line
window: the window writes each line after its prompt, takes a blank
line, cls, reset and exit itself, keeps the lines for the arrow keys,
and keeps its text, history and working directory over a stop and a
start.
The room now holds a command line window per computer for everyone
(TerminalWindow), kept over stops; the protocol carries its lines and
the one way to read them (terminalScreen); the core goes on in the
working directory of the run before (continueIn). The desktop opens
programs in Java's frame, and the command line works with keyboard and
finger: grey on black, the prompt hidden while a command runs, history,
chips for the commands, Ctrl+C. The room's failure is now tested in the
browser.
Fixed: the room handed the engine the wall clock with fractions of a
millisecond, and a ping hung after its first timeout. The engine and
the room count whole milliseconds, as Java's clock does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Measured first: SheetRun now types into FILIUS' DHCP dialog and keeps rows
in its port forwarding dialog. Java's DHCP server keeps its bounds as typed
(".0.0.1" too), never takes an empty gateway or DNS server, hands out the
device's gateway and DNS server unless its own (a home router always its
LAN address, where "manual settings" do nothing), drops a second static
assignment of a MAC address silently and lets "_" through its MAC pattern;
the port forwarding dialog keeps any row, its firmware forwards only the
valid ones. Here a value passes its rule or is not taken, with a word why.
- doc: DHCP server addresses and flags, static assignments (a MAC address
once, ignoring case), port forwarding rows (TCP or UDP, musterPort, the
LAN address); the sheet shows what the server hands out and which rows
Java's firmware forwards (the core's parsePortForwardingRow, exported)
- editor: Java's dialogs as pages of the sheet with a back button, the
three row tables on one RowsEditor (the forwarding table rebuilt on it);
a sheet per device, so an open page closes when another device is chosen
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
packages/editor (React 18, Vite) draws a document as Java FILIUS does: Java's
default hardware icons, the device button of JSidebarButton, cables as
JCablePanel's quadratic curve between the buttons' middles, wireless cables
dashed, documentation items below. The surface moves with mouse and finger
(the pointer captured only once it moved) and zooms with Ctrl and the
wheel, a trackpad pinch, two fingers and the zoom bar; the first view shows
all. German and English catalogs with Java's device names, the dashboard's
colour tokens, the bridge with its origin check. packages/doc gets the
drawing's light view of a document.
Java's Switch.cloud (the Internet in the examples) was lost on reading; it
is now read, kept in the model and written, with FlsOracle and the corpus,
written and exported checks renewed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GnutellaRun adds a line of four peers: a search goes on as a query, each
peer answers with a hit for each matching file and the hits come back the
way the query came; a hit's file is fetched by HTTP on port 6346. The
file type names of the model move from the project export and import into
the core, so that the download uses them too.
Departures from Java: a peer that sent hits keeps its connection, so that
the hits of farther peers come back too (Java closes it and loses them),
and sizes are in characters.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GnutellaRun lets Java's PeerToPeerAnwendung join peers on terminal-net.
The core's gnutella.ts does the same: a ping in Java's text format to
port 6346, a pong back, the ping sent on to the other peers and their
pongs back the way it came, at most the program's number of peers.
Departure from Java: a ping that comes a second time gets no pong and
its connection stays (Java's worker throws on the missing pong).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
FileCases does what the windows of FILIUS' file explorer and text editor
do to a host's files: new folder, delete, copy, cut and paste, rename,
import, save and save as. The core's filePrograms.ts does the same on the
node's file system; the file system renames an entry in its place, as
Java's tree does.
Departures from Java: a folder name that comes to no name makes nothing,
a paste onto a name there already is refused, a rename to a name with a
slash or blanks at its ends is refused and one of the case only is done,
save as and import onto a folder's name are refused.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The account file's writer moves from the export (packages/fls) into the
core, so that the export and the running server write the same. The
server writes the file at its start and whenever its accounts change.
EmailRun now records Java's file as it left it and as Java writes the
accounts it ended with; Java leaves the notice to a local sender out until
its next write.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TerminalRun adds nine stories: traceroute in the LAN, through the router,
unreachable, without a route, to no one, to the host itself, errors and
Ctrl+C; tcpdump on two hosts. traceroute counts failures across hops as
Java does and takes an error that carries its probe; tcpdump tells each
frame's highest capture row with its time of day, at once.
Departures from Java: traceroute tells its summary before the end mark
(Java after it, where its window shows no text) and calls a host at the
20th hop reached. arpsend is not built.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TerminalRun types lines into Java's command line on terminal-net at set
times; the engine types the same lines. New in the engine: an ICMP echo
in a program's thread (the first ICMP message after the send, whatever
it is), Java's route choice, the ARP table with Java's broadcast row, DNS
in a program's thread, ICMP broadcasts of a node's own network, and an
ARP wait that ends after what else happens at that millisecond.
Departures from Java: an echo is answered from the address asked
(RFC 792), network unreachable carries the packet with its sender, ping
names the host and takes only -b before it, arp and netstat list in the
order learned, netstat lists a UDP server's listening socket only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The terminal splits a line as Java's window does and runs ls, cd, pwd,
mkdir, touch, cp, mv, rm, cat, test, help, ipconfig and route on the
node's files, which are the project's own. Its texts are generated from
the bundle Java loads. TerminalCases feeds Java's Terminal without a
network (16 cases, de and en); eight cases name Java bugs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The program sends to its account's SMTP server, one connection for each
recipient, and fetches its mails by POP3; the server now sends mails on
to other domains by MX and returns notices to their senders. The
program's folders are in the snapshot and go back into the project.
EmailClientRun plays six stories with two servers; the mail oracles run
in the language's locale, so Java's notice dates match the core's.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
SMTP on port 25 stores mails for the domain's accounts as Java reads
them anew and tells a local sender of unknown recipients; POP3 on port
110 hands them out. The mailboxes are in the snapshot and go back into
the project. EmailRun plays four stories over the network; MailCases
feeds Java's workers without one (26 cases, de and en). 18 cases name
Java bugs that the core does not take.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The page konfig changes the firewall at run time (into the snapshot, back
into the project through projectWithSimulationChanges). KonfigCases feeds
Java's plug-in directly (25 cases, de and en); RouterWebRun plays two
stories over the network. Ten Java bugs named, not adopted. The web server
answers a method not written GET or POST with 501, as Java does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Routers and home routers start a web server on port 80 with Java's index,
log and routes pages in the project's language (routerPages.ts); the
firewall keeps a log per node that outlives a new start; RIP routes carry
the telling router's public address. TCP counts sequence numbers and
segments in Java's characters, not bytes. RouterWebRun records six stories,
with firewalls now built by ScenarioNetwork. Not taken from Java: its router
server sometimes never listening (a race in setAktiv), and </titel>.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The servers share ServerAnwendung's frame (server.ts); the web server serves
/webserver with virtual hosts, the browser loads a page and its images over
one connection to port 80 (web.ts); HTTP in Java's text form (http.ts).
WebRun records eleven stories on the new web-lan network, with the server's
files; webJava.test.ts plays them card by card. A program's close is a
request, as Java's schliessen. Not taken from Java: /../ serving any file of
the host; an unreadable request closing the connection without a word.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Programs run as generators that yield what they wait for (threads.ts); the
engine resumes them after every event. TCP connect, send, receive, accept
and name resolution exist as such steps; the blocking calls wrap them.
EchoRun records eight stories in Java FILIUS, frames, tables and each
program's window lines; echoJava.test.ts plays them card by card. Not
taken from Java: a connection without a route hangs for ever; a server
switched off leaves its connections' last FIN unanswered.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ExampleRun opens an example as FILIUS opens it (SzenarioVerwaltung.laden
into the window's lists), starts it as the action mode does, plays a
story and writes every card's frames and tables, the calls, the
addresses and the RIP tables' changes. examplesJava.test.ts reads the
same file with this package's reader and runs it in the core
(projectSnapshot, new): routing (internetwork_komplett), DHCP
(einfaches_rechnernetz_komplett), DNS (dns_server), NAT
(nat_with_mapped_port) and firewall (Internet_Komplett_..._DE) match
card by card.
Found and fixed: a host's ipForwardingEnabled, which Java saves and
obeys, was unknown to the reader, the model, the core, the writer, the
export and FlsOracle; and Java's ARP learns only from a request for its
own address or a reply. Named Java bugs: a ping without a gateway throws
a NullPointerException; a home router drops the answer of a second
remote end. Open: Java picks between equally long RIP routes by thread
timing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
captureRows.ts builds Java's rows per frame (Lauscher.datenVorbereiten):
Ethernet; ARP, ICMP or IP; TCP or UDP; the data with Java's protocol
recognition; with addresses and ports or ports only. The comment stays
structured; captureCommentText says it with given words and numbers.
ScenarioNetwork writes Java's finished tables per card (Lauscher.getDaten,
German locale); all runtime runs are recorded again, NatRun has ttlOut.
expectSameCaptures compares both views for every frame. Fixed where the
engine differed: an ARP request's target MAC is the broadcast address,
an ICMP error's identifier and sequence number are 0 (RFC 792), and host
unreachable from a router carries the packet as forwarded. Not taken:
Java's lower-case broadcast address in a reply to 0.0.0.0, and the typo
'Exeeded'. RIP keeps its recordings: Java's staggered router start now
gives a different startup on this machine.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Java's RIPTable.check marks an expired route only at the router's next
beacon, up to some 35 s late. The engine schedules each router's first
expiry, marks the route with 16 hops at that time and sends its table
at once, as after any change (better than Java where Java only samples,
decided 01.10.2026). ripJava's expire story is compared up to the
expiry and then checks that R2 marks the route at its time and R1 a
cable later.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Java's GatewayFirmware.checkNAT and SwitchFirmware.checkSAT look every
second from the start, so an entry goes up to a second late. The engine
expires both exactly at their age: better than Java where Java only
samples, as long as saved files stay compatible (decided 01.10.2026).
The switch did so already; its departure is now named in the code.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The engine keeps Java's packet capture (capture.ts): per network card
every frame it sent (when it leaves on the cable) or received (also one
for another card), once, with time and number; switches log nothing;
frames the packet loss dropped are kept apart; the capture outlives a
stop and a new start, as in Java, until clearCapture. Per port instead
of per MAC, at most 10,000 frames a card.
expectSameCaptures compares each card's capture with Java's list of that
card in the ARP, WLAN, DNS, TCP, NAT, RIP and single-client DHCP runs.
It found that ScenarioNetwork ignored a switch's learning time (Java ran
300 s, the core the scenarios' 5 s); the oracle sets it now, and the TCP
and DHCP runs are recorded again. A core test found that a new start
numbers frames from 1 again and the capture took them as seen.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
WlanRun pings on the new network wlan-lan: over wireless cables, over a
plain cable, and to and from a notebook on the switch's SSID without a
cable. The engine matches Java without a change: it runs the links only.
Swift's wirelessAssociations, which join a host to its SSID's switch at
run time, are not taken. ScenarioNetwork sets wireless cables, wireless
cards and SSIDs; a corpus test pings over the wireless cables of
multicast_server.fls.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The core translates before the IP thread, as Java's NAT threads sit
between card and Ethernet: out, the WAN address and the entry's port (ICMP
keeps its identifier); in, by remote end, port and protocol, else by the
port forwarding, which then makes a dynamic entry; a ping at the WAN
address is answered by the router, packets from outside for the LAN are
dropped, both sides pass the router's firewall, TTL 1 going out gets time
exceeded. Dynamic entries expire at Java's check each second, 300 s times
the delay factor old (55 minutes at 90 %), refreshed only by what comes in.
The home router forwards as a router does.
Measured and not taken: when a LAN endpoint sends to a second remote end,
Java gives it the entry's WAN port but no entry, and the answer stays at
the router; the core adds an entry for each new remote end on that port.
The oracle now records every card of a node (the second card of routers
was missing); the ARP and RIP runs are recorded anew and still match.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
RipRun runs rip-chain's three routers in Java FILIUS, each router's random
set by reflection to the core's ripSeed, and writes the frames, every
change of each RIP table and the calls. converge: the tables fill, a ping
at 45 s. expire: R3 falls silent at 10 s; R2 marks its network 16 hops only
at its own next beacon, R1 passes it on at once, the ping gets ICMP 3/0.
Both match: per sending card the same messages at the same times, every
table through the same states. The first 5 s are compared by the state the
tables reach, since Java's routers start one after the other.
Measured on the JDK: Java's pause divides nextFloat() by 3 in float; the
core does too now (a millisecond off in one draw of about 10,000 before).
Scenarios name their RIP routers in the new field rip.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Where the Swift template departs from Java's filius.software.rip, Java's
code rules: a learned route gets 16 hops at the router's next beacon after
it expired (RIPTable.check), not when it expires; messages count from any
source port; each router draws its pauses from a random of its own (Java:
an unseeded Random per router), which the core seeds from the scenario and
the router's id (ripSeed), so that the Java oracle can give it the same.
A route of hop count 0 to each own network, the first beacon after a
second, then every 25.2 to 35.2 s; a message per interface without the
routes through it; a new route or a longer one from the same router sends
at once; forwarding by the table alone, fewest hops before longest mask.
Messages are read with Java's parser (javaIpCheck, javaParseInt, moved from
fls to core for it).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The resolver runs on the events now (a query, its dnsQueryTimeout, its
answer at the socket); the blocking resolveDns waits for it, and a
recursive server uses it: its own records, else the zone's name server
from its records, else its own DNS server, else a referral
(DNSServer.answerWithRemoteData).
DnsRun plays five stories on the new dns-recursive network. Measured and
not taken: a server that is its zone's name server asks itself, which asks
itself again, without end (the client times out, then every level floods
its own DNS server); Java guards only the step through its own DNS server.
The core asks itself in neither step. That story is compared up to the
client's query (JAVA_UNTIL).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every packet that comes in goes through its node's IP thread (UDP, TCP) or
ICMP thread (ICMP), as Java's IPThread and ICMPThread: deliver, answer an
echo request or forward, one after the other; while one waits for ARP, the
next wait behind it. A packet a node sends to itself is delivered at once,
as Java's IP.senden does. The transport send threads and these are one
mechanism now (handToProtocolThread).
ArpRun has three stories through the router (routedUdpBehind,
routedPingBesideUdp, routedPingBehindPing); the oracle writes ICMP frames
with type, code, addresses and sequence number. All seven ArpRun stories
match Java FILIUS.
Measured and not taken: Java's ARP wakes one waiter with notify, not the
one whose answer came; the core wakes all that wait for the address.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TcpRun loses exactly one frame (loseFirst: the card's output buffer drops
it; the sender's packet log still shows it, as with setDrop) in three new
stories: lostSynAck, lostFin, lostLastFin.
Measured: Java never repeats a SYN+ACK or a FIN. The server gives a
handshake up after one timeout (3 RTT) and closes with a FIN from SYN_RCVD
(its boundary check sometimes lets it wait twice; it means once). After its
FIN a socket waits at most five times, each wait ended by a segment or the
timeout; then it is gone. The core does the same (TCP_CLOSING_WAITS, with
TIME_WAIT in the loop) and shows CLOSED where Java leaves the state
standing (a named Java bug).
One departure: a repeated SYN in SYN_RCVD gets the SYN+ACK once more and
the wait starts anew; Java closes there, but its client never repeats the
SYN. lostSynAck is compared up to the second SYN (JAVA_UNTIL); the rest is
an engine test.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A send hands its frame to the card and returns, as in Java; the frames
travel while the clock advances. Swift's send ran the network until every
frame had arrived; the option swiftSends keeps that for the Swift
comparison. Waiting calls (TCP connect, acknowledgement) now run event by
event until the answer instead of to the timeout.
UDP and TCP send through a thread per node and protocol (Java's
TransportProtokoll): one segment after the other; while IP waits for ARP,
the next send blocks and a segment from an event queues behind it. Senders
that wait take their turns in the order they came (a fair lock; Java's
leaves the order to chance).
All four ArpRun stories now match Java FILIUS, including when each call
returns. Tests let frames arrive explicitly with letFramesArrive.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An ARP lookup asks, waits a round trip, asks once more and gives up after
another round trip; then host unreachable, observed at the sender itself or
sent back by the router (for ICMP only for an echo request). A waiting
sender wakes after what else happens at that time. A ping then waits a
round trip for its answer, counted from when it left. Swift asks twice at
once while events run and lets the packet wait without end when the
address exists anywhere in the network; the option swiftArp keeps that for
the Swift comparison, and three scenarios are compared with Swift until the
trace where ARP fails or a ping goes unanswered.
arpJava.test.ts compares ArpRun: both pings match; the two UDP stories wait
for Java's transport thread (next step). DNS without a server now matches
Java too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DNS follows Java's DNSServer and Resolver, not Swift's typed format:
Java's text messages (dns.ts), a server on port 53 answering from its
records (no question in the answer, RCODE 0, a referral with NS and A in
the answer section), an iterative resolver that waits five round trips per
query and remembers only the servers it was referred to. Recursion throws
NotPortedError until the next step. The snapshot carries the DNS servers
switched on. receiveUdp returns as soon as a datagram is there, as Java's
empfangen does.
dnsJava.test.ts compares DnsRun's five lookups; four match frame by frame,
the missing server waits for Java's ARP (named as not yet like Java).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Java's SimplexVerbindung sleeps a cable's delay with the port's buffer
locked: frames on one cable leave one after the other, and a frame shows
in the packet log when it leaves. The engine now queues a frame in the
output buffer until the one before has arrived (ethernetLinkDeparture).
Swift sends a cable's frames at once; the option swiftCables keeps that for
the Swift comparison, which goes on checking everything else trace by
trace.
With the queue, the Java comparisons match within 120 ms (was 200); a
state of the TCP stories within 250 ms, where the engine's calls block
one after the other.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TCP follows Java's TCPSocket where it differs from Swift: each attempt and
TIME_WAIT wait three round trips that follow the speed (8250 ms at 90 %),
counted from when the segment is handed down, before ARP; the next segment
ends TIME_WAIT; the other end closes by itself on FIN; data no one
acknowledges closes with FIN; the accepted socket has the listener's
initial sequence number, and the listener takes the next.
Not taken from Java: three SYNs, not one (Java's retry loop ends after
the first timeout), named in tcpJava.test.ts as JAVA_BUGS.
tcpJava.test.ts compares TcpRun's three stories frame by frame and state
by state; the Swift comparison only checks that the TCP scenarios differ.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The client follows Java's DHCPClient instead of Swift: one DISCOVER or
REQUEST per attempt, an answer that does not fit is read and dropped, the
ARP check waits a round trip (an answer for the address ends it at once,
then DECLINE), a timeout repeats the state's message, and the fifth turns
to a link-local address 169.254.1.x/16 (ARP probe, ARP reply to all); no
old address is given back. DHCP round trip times follow the speed as in
Java (50 cables, 2750 ms at 90 %; Swift: fixed 2500 ms).
Not taken from Java: the client gives up after five occupied link-local
addresses (Java searches on without end), and an ARP packet for another
address does not end the check.
dhcpJava.test.ts compares the four DHCP scenarios with DhcpRun's Java
runs; the Swift comparison only checks that they differ.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
